Home › Groups › Discussions › Watching for Flu Outbreak Globally - Assistance in Term Translation Needed

Discussions

User login

Watching for Flu Outbreak Globally - Assistance in Term Translation Needed

Submitted by John Geraci on 25 April 2009 - 12:20pm to Discussions.

Acting on a request from some health researchers in the U.S., we've been trying for the last 24 hours to improve SickCity's algorithms to be able to track new developments of the swine flu outbreak.

After initially being flooded with tweets about the swine flu *media event* as opposed to actual flu outbreaks, the tool has now gotten to be a lot stronger - we're now able to ignore a large percentage of tweets that are about the media event, to focus on tweets about actual sickness. We're also tracking better terms - for instance instead of 'headcold', we're tracking 'fever'. We still have to make it stronger, however - among other things we have to get better at sniffing out false positives (e.g. "spring fever" should not trigger alert for "fever"). We also still have to flush out all of the false positives we have accumulated in the past 24 hours - the site right now looks like a gigantic flu outbreak happened all over the world simultaneously. That will be fixed some time today.

What we really need right now is translation of terms into other languages. All of the countries in SickCity are tracking terms in english ("fever", "sore throat", "flu", etc). If anyone can help translate these terms and others into other languages, that would be a big help. We've set up a wiki page where you can do this. Should only take one minute to do.

Anyone else wanting to pitch in to make SickCity more accurate (especially anyone good at term extraction and elimination of false positives) please join the SickCity Development Group and post a note.

25 Apr13:51

you can also just email

By John Geraci

you can also just email translations directly to me at geraci...gmaildotcom if that's easier and I'll send them to the developers.

26 Apr15:48

Translation for some terms to argentinian spanish

By marianogoren

Hi folks.
I'm new in DIYC community, and this is my first post.

I've added translation for some terms to argentinian spanish. In most topics, I've tried to add adjectives and expressions aswell. I understand that if we want to track accurately, we must watch out for zone language distinctions.

BTW I'm opening the argentinian first group, DIYCity Buenos Aires.

Cheers,

Mariano A. Goren
www.marianogoren.com.ar

28 Apr15:07

translation of terms, etc.

By John Geraci

Thanks Mariano, and all of the others who have emailed translations in to us.

With this having become such an enormous story around the world over the past few days, the biggest challenge right now is to filter out all of the messages from people just registering their concern, or interest, in the story, in order to hone in on the messages that are really about people being sick.

To do that, the SickCity developers have set up a page where people can add to a list of "false positive" terms that should be ignored. This is helping get the noise out of the system. You can see this, and add to it if you want, at http://sickcity.org/badwords.

Their next step is going to be to allow visitors to remove noisy messages from the system themselves - each visitor can be an editor of the data essentially. This seems like the only way to get data that is truly free of noise.

We want to also allow people to add terms in languages other than English on the site itself, to get better term search going on in other countries. Hopefully will get to that soon.

28 Apr15:50

Re: translation of terms, etc.

By achanda

Hey everyone,Just joined diycity this morning. I hope I can make some useful contribution here. I had a question/suggestion about collecting false positive terms for monitoring online chatter.

It would be useful if we knew what words Sickcity was monitoring themselves. Then it would give us an idea about what misleading phrases could be built out of those words and we could add them to the list of false positives. Will this be helpful?

Thanks.amritaOn Tue, Apr 28, 2009 at 11:07 AM, John Geraci <discussions@diycity.org> wrote:

comment 'translation of terms, etc.' by John Geraci

Thanks Mariano, and all of the others who have emailed translations in
to us.  

With this having become such an enormous story around the world over
the past few days, the biggest challenge right now is to filter out all
of the messages from people just registering their concern, or interest,
in the story, in order to hone in on the messages that are really about
people being sick.  

To do that, the SickCity developers have set up a page where people can
add to a list of "false positive" terms that should be ignored.  This is
helping get the noise out of the system.  You can see this, and add to
it if you want, at http://sickcity.org/badwords [2].

Their next step is going to be to allow visitors to remove noisy
messages from the system themselves - each visitor can be an editor of
the data essentially.  This seems like the only way to get data that is
truly free of noise.

We want to also allow people to add terms in languages other than
English on the site itself, to get better term search going on in other
countries.  Hopefully will get to that soon.

[2] http://sickcity.org/badwords

View original: http://diycity.org/discussions/watching-flu-outbreak-globally-assistance-term-translation-needed#comment-391

Post reply: http://diycity.org/comment/reply/468/391, or you can post a new post by e-mail: discussions@diycity.org

Note that if you reply to this mail delete the quoted text and do not modify the subject field or the message will not reach it's destiny.
--
You are subscribed from the group 'DIYcity Main Group' at DIYcity.
To manage your subscription, visit http://diycity.org/og/manage/16.
-- A uniforming globalization throws up a contradiction between efficiency, driven by interconnectedness, and resilience (to collapse), which needs diversity and some degree of separation.

- "Our panarchic future", Thomas Homer-Dixonhttp://www.worldwatch.org/node/6008